Can your model explain where it came from?

One of these models can show you its origin.
Where did the other come from?

hash only artifact only
roar runtime lineage captured
LLM
nanochat
sha256: 7fa2d9c4a3b2e1f0d8c591aa...
  • dataset version unknown
  • preprocessing unavailable
  • cannot reproduce exactly
LLM
nanochat
glaas://artifact/2a38451304ce1fc6...
  • dataset version linked
  • preprocessing traceable
  • reproducible from source runs
view lineage →

roar captures what actually ran. GLaaS lets you look it up by hash.

try it yourself

Copy a hash from the right column above and paste it here.

give your next artifact execution history
$uv tool install roar-cli
or: pipx install roar-cli  ·  pip install roar-cli

No account required·~30 seconds·no code changes

Run this on your next training job.

Provenance captured during execution — not reconstructed later from logs.

Supports traceability requirements under emerging AI regulation.